The goal of the project is to develop a rapid, automated, non-invasive method for comprehensive evaluation of left ventricular contractile performance utilizing a 3-dimensional time varying reconstruction of 2-dimensional echocardiograms. This reconstruction involves realignment of five echocardiographic cross sections on a single longitudinal axis view of the ventricle. A framework for this realignment utilizing 61 specific points from both the epicardium and endocardium has been developed. The computer software for realignment of the cross sections of this framework and for the calculation of quantitative indices of left ventricular performance have been completed in a preliminary form. Theoretical work with these programs strongly suggests that the largest source of error involves correct identification of the epicardial and endocardial boundaries. Several concurrent investigations are proposed for the three years of research. The first, based on the preliminary work, involves investigation to delineate the best resolution of boundaries of which the 2-dimensional echocardiograph is capable. Thiw will be done through in vitro studies of grids, animate and inanimate artificial objects, and animate and inanimate isolated hearts. Second, using these results, calibration curves will be determined for the best estimate of numerous indices of cardiac performance and the ability of the 3-D reconstruction to accurately define segmental wall motion abnormalities will be explored in isolated heart preparations. Third, in vivo clinical studies will be performed and 3-D echocardiographic results compared to simultaneous flow velocity indices obtained during cardiac catheterization. Concurrent with the above investigations, on line analog to digital conversion of the real time signal will be accomplished and techniques for rapid automated border recognition will be developed.